Frequency domain-based auto tuning controller for plants with uncertainties using well known tuning formulas

Abstract

Adaptive control which adequately adjusts controller gains according to the changes in plants, has become attractive in recent years. The controller proposed in this paper is tuned automatically with various tuning formulas based on the results of frequency domain system identification for the plant. The controller first estimates the frequency response of the plant using FFT. The controller gains are automatically tuned so as to minimise the error between the open loop frequency response of the reference model and that of the actual system at a few frequency points. For the three example processes, reference models are derived. The frequency responses of the reference models and that of the actual processes are obtained. The controller gains are determined by applying the least squares algorithm. The responses of the plants are verified in time domain and frequency domain after tuning the I-PD controller.